Learning Forum

[C B PUNITH GOWDA]

Respected member,

   I have been trying to simulate a model in Ansys FLUENT for a random wave condition. But even after the simulation is completed, there is no wave pattern visible in the contours window. Whereas for the same model, we were able to get results for regular wave conditions, but for random waves we are not getting to see any wave patterns. Thus, could you please guide us technically on how to go about for random wave analysis.

   I have inserted images on what we have input, kindly let us know of the changes to be incorporated. 

   Looking forward to your response.  Thank you.

             

[Rob]

How well resolved is the mesh? 

[C B PUNITH GOWDA]

Dear Rob,

      Please find below the attached images wherein the mesh sizing and corresponding information is input. Kindly guide me through this and let me know of any corrections.

[Rob]

So, not 10-20 cells in the wave amplitude? 

[C B PUNITH GOWDA]

Dear Rob,

   The range of wave heights used in our work is 10 cm to 16 cm. Thus we get 2-3 cells in the wave amplitude. But for the same setup, we are able to obtain the results for regular wave condition. When it comes to random wave, we are a little clueless. Is it recommended to maintain 10-20 cells in the wave amplitude.

[Rob]

Typically you need a higher mesh resolution. Random waves may also take longer to run into the domain, how long did you run for?

[C B PUNITH GOWDA]

Dear Rob,

   We have run the simulation for a period of 22 sec (i.e 2200 timesteps, each time step=0.01 sec). It takes around 26 hours for completing these time steps. 

   The system configuration is; i7 10th gen, 8 core processor-16 processes.   

   Can you please suggest the suitable mesh resolution that might be necessary for this condition. And if any improvements needed.

Thank you.

[Rob]

You need 10-20 cells over the amplitude to resolve the free surface and to keep the aspect ratio low (2-3 max) around the free surface. If the chip has 8 cores, and hyperthreading just run on 8 parallel, hyperthreading and parallel aren't an efficient combination.